Merge branch 'vera' into main

main.py

@@ -3,6 +3,9 @@ import logging
 from injector import Injector,inject
 from src.configuration import LogConf, singleton
 
+import argparse
+
+
 @singleton
 class Main():
 
@@ -13,7 +16,7 @@ class Main():
     def run(self):
         logger = logging.getLogger(__name__)
         logger.info("jarvis-models start", extra={"version": "0.0.1"})
-        uvicorn.run("server:app", host="0.0.0.0", port=8000, log_level="info")
+        uvicorn.run("server:app", host="0.0.0.0", port=8000, log_level="info",reload = True)
 
 if __name__ == "__main__":
     injector = Injector()

sample/chroma_client1.py

@@ -70,7 +70,7 @@ def get_all_files(folder_path):
 
 
 # 加载文档和拆分文档
-loader = TextLoader("/home/administrator/Workspace/jarvis-models/sample/RAG_zh.txt")
+loader = TextLoader("/home/gpu/Workspace/jarvis-models/sample/RAG_zh.txt")
 
 documents = loader.load()
 
@@ -84,11 +84,11 @@ ids = ["20240521_store"+str(i) for i in range(len(docs))]
 
 
 # 加载embedding模型和chroma server
-embedding_model = SentenceTransformerEmbeddings(model_name='/home/administrator/Workspace/Models/BAAI/bge-large-zh-v1.5', model_kwargs={"device": "cuda"})
-client = chromadb.HttpClient(host='172.16.5.8', port=7000)
+embedding_model = SentenceTransformerEmbeddings(model_name='/home/gpu/Workspace/Models/BAAI/bge-large-zh-v1.5', model_kwargs={"device": "cuda"})
+client = chromadb.HttpClient(host='10.6.81.119', port=7000)
 
 id = "g2e"
-client.delete_collection(id)
+#client.delete_collection(id)
 collection_number = client.get_or_create_collection(id).count()
 print("collection_number",collection_number)
 start_time2 = time.time()
@@ -106,8 +106,8 @@ print("collection_number",collection_number)
 
 # # chroma 召回
 # from chromadb.utils import embedding_functions
-# embedding_model = embedding_functions.SentenceTransformerEmbeddingFunction(model_name="/home/administrator/Workspace/Models/BAAI/bge-large-zh-v1.5", device = "cuda")
-# client = chromadb.HttpClient(host='172.16.5.8', port=7000)
+# embedding_model = embedding_functions.SentenceTransformerEmbeddingFunction(model_name="/home/gpu/Workspace/Models/BAAI/bge-large-zh-v1.5", device = "cuda")
+# client = chromadb.HttpClient(host='10.6.81.119', port=7000)
 # collection = client.get_collection("g2e", embedding_function=embedding_model)
 
 # print(collection.count())
@@ -152,7 +152,7 @@ print("collection_number",collection_number)
 #     'Content-Type': 'application/json',
 #     'Authorization': "Bearer " + key
 # }
-# url = "http://172.16.5.8:23333/v1/chat/completions"
+# url = "http://10.6.81.119:23333/v1/chat/completions"
 
 # fastchat_response = requests.post(url, json=chat_inputs, headers=header)
 # # print(fastchat_response.json())

sample/chroma_client_en.py

@@ -70,7 +70,7 @@ def get_all_files(folder_path):
 
 
 # 加载文档和拆分文档
-loader = TextLoader("/home/administrator/Workspace/jarvis-models/sample/RAG_en.txt")
+loader = TextLoader("/home/gpu/Workspace/jarvis-models/sample/RAG_en.txt")
 
 documents = loader.load()
 
@@ -84,8 +84,8 @@ ids = ["20240521_store"+str(i) for i in range(len(docs))]
 
 
 # 加载embedding模型和chroma server
-embedding_model = SentenceTransformerEmbeddings(model_name='/home/administrator/Workspace/Models/BAAI/bge-large-en-v1.5', model_kwargs={"device": "cuda"})
-client = chromadb.HttpClient(host='172.16.5.8', port=7000)
+embedding_model = SentenceTransformerEmbeddings(model_name='/home/gpu/Workspace/Models/BAAI/bge-small-en-v1.5', model_kwargs={"device": "cuda"})
+client = chromadb.HttpClient(host='10.6.81.119', port=7000)
 
 id = "g2e_english"
 client.delete_collection(id)
@@ -106,8 +106,8 @@ print("collection_number",collection_number)
 
 # # chroma 召回
 # from chromadb.utils import embedding_functions
-# embedding_model = embedding_functions.SentenceTransformerEmbeddingFunction(model_name="/home/administrator/Workspace/Models/BAAI/bge-large-zh-v1.5", device = "cuda")
-# client = chromadb.HttpClient(host='172.16.5.8', port=7000)
+# embedding_model = embedding_functions.SentenceTransformerEmbeddingFunction(model_name="/home/gpu/Workspace/Models/BAAI/bge-large-zh-v1.5", device = "cuda")
+# client = chromadb.HttpClient(host='10.6.81.119', port=7000)
 # collection = client.get_collection("g2e", embedding_function=embedding_model)
 
 # print(collection.count())
@@ -152,7 +152,7 @@ print("collection_number",collection_number)
 #     'Content-Type': 'application/json',
 #     'Authorization': "Bearer " + key
 # }
-# url = "http://172.16.5.8:23333/v1/chat/completions"
+# url = "http://10.6.81.119:23333/v1/chat/completions"
 
 # fastchat_response = requests.post(url, json=chat_inputs, headers=header)
 # # print(fastchat_response.json())

sample/chroma_client_query.py

@@ -66,7 +66,7 @@ def get_all_files(folder_path):
 
 
 # 加载文档和拆分文档
-# loader = TextLoader("/home/administrator/Workspace/jarvis-models/sample/RAG_zh.txt")
+# loader = TextLoader("/home/gpu/Workspace/jarvis-models/sample/RAG_zh.txt")
 
 # documents = loader.load()
 
@@ -80,8 +80,8 @@ def get_all_files(folder_path):
 
 
 # # 加载embedding模型和chroma server
-# embedding_model = SentenceTransformerEmbeddings(model_name='/home/administrator/Workspace/Models/BAAI/bge-large-zh-v1.5', model_kwargs={"device": "cuda"})
-# client = chromadb.HttpClient(host='172.16.5.8', port=7000)
+# embedding_model = SentenceTransformerEmbeddings(model_name='/home/gpu/Workspace/Models/BAAI/bge-large-zh-v1.5', model_kwargs={"device": "cuda"})
+# client = chromadb.HttpClient(host='10.6.81.119', port=7000)
 
 # id = "g2e"
 # client.delete_collection(id)
@@ -102,8 +102,8 @@ def get_all_files(folder_path):
 
 # chroma 召回
 from chromadb.utils import embedding_functions
-embedding_model = embedding_functions.SentenceTransformerEmbeddingFunction(model_name="/home/administrator/Workspace/Models/BAAI/bge-large-zh-v1.5", device = "cuda")
-client = chromadb.HttpClient(host='172.16.5.8', port=7000)
+embedding_model = embedding_functions.SentenceTransformerEmbeddingFunction(model_name="/home/gpu/Workspace/Models/BAAI/bge-large-zh-v1.5", device = "cuda")
+client = chromadb.HttpClient(host='10.6.81.119', port=7000)
 collection = client.get_collection("g2e", embedding_function=embedding_model)
 
 print(collection.count())
@@ -148,7 +148,7 @@ print("time: ", time.time() - start_time)
 #     'Content-Type': 'application/json',
 #     'Authorization': "Bearer " + key
 # }
-# url = "http://172.16.5.8:23333/v1/chat/completions"
+# url = "http://10.6.81.119:23333/v1/chat/completions"
 
 # fastchat_response = requests.post(url, json=chat_inputs, headers=header)
 # # print(fastchat_response.json())

src/blackbox/asr.py

@@ -6,26 +6,109 @@ from fastapi.responses import JSONResponse
 
 from ..asr.rapid_paraformer.utils import read_yaml
 from ..asr.rapid_paraformer import RapidParaformer
+
+from funasr import AutoModel
+from funasr.utils.postprocess_utils import rich_transcription_postprocess
 from .blackbox import Blackbox
 from injector import singleton, inject
 
+import tempfile
+
+import json
+import os
+from ..configuration import SenseVoiceConf
+
+from ..log.logging_time import logging_time
+import logging
+logger = logging.getLogger(__name__)
+
 @singleton
 class ASR(Blackbox):
+    mode: str
+    url: str
+    speed: int
+    device: str
+    language: str
+    speaker: str
+
+    @logging_time(logger=logger)
+    def model_init(self, sensevoice_config: SenseVoiceConf) -> None:
+
+        config = read_yaml(".env.yaml")
+        self.paraformer = RapidParaformer(config)
+
+        model_dir = "/home/gpu/Workspace/Models/SenseVoice/SenseVoiceSmall"
+
+        self.speed = sensevoice_config.speed
+        self.device = sensevoice_config.device
+        self.language = sensevoice_config.language
+        self.speaker = sensevoice_config.speaker
+        self.device = sensevoice_config.device
+        self.url = ''
+        self.mode = sensevoice_config.mode
+        self.asr = None
+        self.speaker_ids = None
+        # os.environ['CUDA_VISIBLE_DEVICES'] = str(sensevoice_config.device)
+        if self.mode == 'local':
+            self.asr = AutoModel(
+            model=model_dir,
+            trust_remote_code=True,
+            remote_code= "/home/gpu/Workspace/SenseVoice/model.py",
+            vad_model="fsmn-vad",
+            vad_kwargs={"max_single_segment_time": 30000},
+            device=self.device,
+        )   
+            
+        else:
+            self.url = sensevoice_config.url
+        logging.info('#### Initializing SenseVoiceASR Service in cuda:' + sensevoice_config.device + ' mode...')
 
     @inject
-    def __init__(self,path = ".env.yaml") -> None:
-        config = read_yaml(path)
-        self.paraformer = RapidParaformer(config)
-    
+    def __init__(self, sensevoice_config: SenseVoiceConf, settings: dict) -> None:
+        self.model_init(sensevoice_config)
+ 
     def __call__(self, *args, **kwargs):
         return self.processing(*args, **kwargs)
     
-    async def processing(self,  *args, **kwargs):
+    async def processing(self,  *args, settings: dict):
+
+        print("\nChat Settings: ", settings)
+        if settings is None:
+            settings = {}
+        user_model_name = settings.get("asr_model_name")
+        print(f"asr_model_name: {user_model_name}")
         data = args[0]
-        results = self.paraformer([BytesIO(data)])
-        if len(results) == 0:
-            return None
-        return results[0]
+
+        if user_model_name == 'sensevoice' or ['sensevoice']:        
+            # 创建一个临时文件
+            with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as temp_audio_file:
+                temp_audio_file.write(data)
+                temp_audio_path = temp_audio_file.name
+            res = self.asr.generate(
+                input=temp_audio_path,
+                cache={},
+                language="auto",  # "zh", "en", "yue", "ja", "ko", "nospeech"
+                use_itn=True,
+                batch_size_s=60,
+                merge_vad=True,  #
+                merge_length_s=15,
+            )        
+            # results = self.paraformer([BytesIO(data)])
+            results = rich_transcription_postprocess(res[0]["text"])
+            os.remove(temp_audio_path)
+            if len(results) == 0:
+                return None
+            return results
+        
+        elif user_model_name == 'funasr' or ['funasr']:
+            
+            return results
+        
+        else:
+            results = self.paraformer([BytesIO(data)])
+            if len(results) == 0:
+                return None
+            return results[0]       
 
     def valid(self, data: any) -> bool:
         if isinstance(data, bytes):
@@ -34,12 +117,20 @@ class ASR(Blackbox):
     
     async def fast_api_handler(self, request: Request) -> Response:
         data = (await request.form()).get("audio")
+        setting: dict = (await request.form()).get("settings")
+        
+        if isinstance(setting, str):
+            try:
+                setting = json.loads(setting)  # 尝试将字符串转换为字典
+            except json.JSONDecodeError:
+                return JSONResponse(content={"error": "Invalid settings format"}, status_code=status.HTTP_400_BAD_REQUEST)
+        
         if data is None:
-            self.logger.warn("asr bag request","type", "fast_api_handler", "api", "asr")
+            # self.logger.warn("asr bag request","type", "fast_api_handler", "api", "asr")
             return JSONResponse(content={"error": "data is required"}, status_code=status.HTTP_400_BAD_REQUEST)
         d = await data.read()
         try:
-            txt = await self.processing(d)
+            txt = await self.processing(d, settings=setting)
         except ValueError as e:
             return JSONResponse(content={"error": str(e)}, status_code=status.HTTP_400_BAD_REQUEST)
         return JSONResponse(content={"text": txt}, status_code=status.HTTP_200_OK)

src/blackbox/asr_bak.py

@@ -0,0 +1,45 @@
+from io import BytesIO
+from typing import Any, Coroutine
+
+from fastapi import Request, Response, status
+from fastapi.responses import JSONResponse
+
+from ..asr.rapid_paraformer.utils import read_yaml
+from ..asr.rapid_paraformer import RapidParaformer
+from .blackbox import Blackbox
+from injector import singleton, inject
+
+@singleton
+class ASR(Blackbox):
+
+    @inject
+    def __init__(self,path = ".env.yaml") -> None:
+        config = read_yaml(path)
+        self.paraformer = RapidParaformer(config)
+    
+    def __call__(self, *args, **kwargs):
+        return self.processing(*args, **kwargs)
+    
+    async def processing(self,  *args, **kwargs):
+        data = args[0]
+        results = self.paraformer([BytesIO(data)])
+        if len(results) == 0:
+            return None
+        return results[0]
+
+    def valid(self, data: any) -> bool:
+        if isinstance(data, bytes):
+            return True
+        return False
+    
+    async def fast_api_handler(self, request: Request) -> Response:
+        data = (await request.form()).get("audio")
+        if data is None:
+            # self.logger.warn("asr bag request","type", "fast_api_handler", "api", "asr")
+            return JSONResponse(content={"error": "data is required"}, status_code=status.HTTP_400_BAD_REQUEST)
+        d = await data.read()
+        try:
+            txt = await self.processing(d)
+        except ValueError as e:
+            return JSONResponse(content={"error": str(e)}, status_code=status.HTTP_400_BAD_REQUEST)
+        return JSONResponse(content={"text": txt}, status_code=status.HTTP_200_OK)

src/blackbox/asrsensevoice.py

@@ -0,0 +1,103 @@
+from io import BytesIO
+from typing import Any, Coroutine
+
+from fastapi import Request, Response, status
+from fastapi.responses import JSONResponse
+
+from funasr import AutoModel
+from funasr.utils.postprocess_utils import rich_transcription_postprocess
+from .blackbox import Blackbox
+from injector import singleton, inject
+
+import tempfile
+
+import os
+from ..configuration import SenseVoiceConf
+
+from ..log.logging_time import logging_time
+import logging
+logger = logging.getLogger(__name__)
+
+@singleton
+class ASR(Blackbox):
+    mode: str
+    url: str
+    speed: int
+    device: str
+    language: str
+    speaker: str
+
+    @logging_time(logger=logger)
+    def model_init(self, sensevoice_config: SenseVoiceConf) -> None:
+
+        model_dir = "/home/gpu/Workspace/Models/SenseVoice/SenseVoiceSmall"
+
+        self.speed = sensevoice_config.speed
+        self.device = sensevoice_config.device
+        self.language = sensevoice_config.language
+        self.speaker = sensevoice_config.speaker
+        self.device = sensevoice_config.device
+        self.url = ''
+        self.mode = sensevoice_config.mode
+        self.asr = None
+        self.speaker_ids = None
+        os.environ['CUDA_VISIBLE_DEVICES'] = str(sensevoice_config.device)
+        if self.mode == 'local':
+            self.asr = AutoModel(
+            model=model_dir,
+            trust_remote_code=True,
+            remote_code= "/home/gpu/Workspace/SenseVoice/model.py",
+            vad_model="fsmn-vad",
+            vad_kwargs={"max_single_segment_time": 30000},
+            device="cuda:0",
+        )   
+            
+        else:
+            self.url = sensevoice_config.url
+        logging.info('#### Initializing SenseVoiceASR Service in cuda:' + str(sensevoice_config.device) + ' mode...')
+
+    @inject
+    def __init__(self, sensevoice_config: SenseVoiceConf) -> None:
+        self.model_init(sensevoice_config)
+ 
+    def __call__(self, *args, **kwargs):
+        return self.processing(*args, **kwargs)
+    
+    async def processing(self,  *args, **kwargs):
+        data = args[0]
+        # 创建一个临时文件
+        with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as temp_audio_file:
+            temp_audio_file.write(data)
+            temp_audio_path = temp_audio_file.name
+        res = self.asr.generate(
+            input=temp_audio_path,
+            cache={},
+            language="auto",  # "zh", "en", "yue", "ja", "ko", "nospeech"
+            use_itn=True,
+            batch_size_s=60,
+            merge_vad=True,  #
+            merge_length_s=15,
+        )        
+        # results = self.paraformer([BytesIO(data)])
+        results = rich_transcription_postprocess(res[0]["text"])
+        os.remove(temp_audio_path)
+        if len(results) == 0:
+            return None
+        return results
+
+    def valid(self, data: any) -> bool:
+        if isinstance(data, bytes):
+            return True
+        return False
+    
+    async def fast_api_handler(self, request: Request) -> Response:
+        data = (await request.form()).get("audio")
+        if data is None:
+            # self.logger.warn("asr bag request","type", "fast_api_handler", "api", "asr")
+            return JSONResponse(content={"error": "data is required"}, status_code=status.HTTP_400_BAD_REQUEST)
+        d = await data.read()
+        try:
+            txt = await self.processing(d)
+        except ValueError as e:
+            return JSONResponse(content={"error": str(e)}, status_code=status.HTTP_400_BAD_REQUEST)
+        return JSONResponse(content={"text": txt}, status_code=status.HTTP_200_OK)

src/blackbox/audio_chat.py

@@ -3,7 +3,7 @@ from fastapi import Request, Response,status
 from fastapi.responses import JSONResponse
 from injector import inject, singleton
 
-from .asr import ASR
+from .asrsensevoice import ASR
 from .tesou import Tesou
 from .tts import TTS
 

src/blackbox/blackbox_factory.py

@@ -47,15 +47,15 @@ def vlms_loader():
     from .vlms import VLMS
     return Injector().get(VLMS)
 
-@model_loader(lazy=blackboxConf.lazyloading)
-def melotts_loader():
-    from .melotts import MeloTTS
-    return Injector().get(MeloTTS)
+# @model_loader(lazy=blackboxConf.lazyloading)
+# def melotts_loader():
+#     from .melotts import MeloTTS
+#     return Injector().get(MeloTTS)
 
-@model_loader(lazy=blackboxConf.lazyloading)
-def cosyvoicetts_loader():
-    from .cosyvoicetts import CosyVoiceTTS
-    return Injector().get(CosyVoiceTTS)
+# @model_loader(lazy=blackboxConf.lazyloading)
+# def cosyvoicetts_loader():
+#     from .cosyvoicetts import CosyVoiceTTS
+#     return Injector().get(CosyVoiceTTS)
 
 @model_loader(lazy=blackboxConf.lazyloading)
 def tts_loader():
@@ -141,11 +141,11 @@ class BlackboxFactory:
         self.models["chroma_query"] = chroma_query_loader
         self.models["chroma_upsert"] = chroma_upsert_loader
         self.models["chroma_chat"] = chroma_chat_loader
-        self.models["melotts"] = melotts_loader
+        # self.models["melotts"] = melotts_loader
         self.models["vlms"] = vlms_loader
         self.models["chat"] = chat_loader
         self.models["chat_llama"] = chat_llama_loader
-        self.models["cosyvoicetts"] = cosyvoicetts_loader
+        # self.models["cosyvoicetts"] = cosyvoicetts_loader
         self.models["workflow"] = workflow_loader
         self.models["sum"] = sum_loader
 

src/blackbox/chat.py

@@ -92,7 +92,7 @@ class Chat(Blackbox):
             #user_presence_penalty = 0.8
  
         if user_model_url is None or user_model_url.isspace() or user_model_url == "":
-            user_model_url = "http://172.16.5.8:23333/v1/chat/completions"
+            user_model_url = "http://10.6.81.119:23333/v1/chat/completions"
 
         if user_model_key is None or user_model_key.isspace() or user_model_key == "":
             user_model_key = "YOUR_API_KEY"

src/blackbox/chroma_query.py

@@ -20,9 +20,9 @@ class ChromaQuery(Blackbox):
     def __init__(self, *args, **kwargs) -> None:
         # config = read_yaml(args[0])
         # load chromadb and embedding model
-        self.embedding_model_1 = embedding_functions.SentenceTransformerEmbeddingFunction(model_name="/home/administrator/Workspace/Models/BAAI/bge-large-zh-v1.5", device = "cuda")
-        self.embedding_model_2 = embedding_functions.SentenceTransformerEmbeddingFunction(model_name="/home/administrator/Workspace/Models/BAAI/bge-large-en-v1.5", device = "cuda")
-        self.client_1 = chromadb.HttpClient(host='172.16.5.8', port=7000)
+        self.embedding_model_1 = embedding_functions.SentenceTransformerEmbeddingFunction(model_name="/home/gpu/Workspace/Models/BAAI/bge-large-zh-v1.5", device = "cuda:1")
+        self.embedding_model_2 = embedding_functions.SentenceTransformerEmbeddingFunction(model_name="/home/gpu/Workspace/Models/BAAI/bge-small-en-v1.5", device = "cuda:1")
+        self.client_1 = chromadb.HttpClient(host='10.6.81.119', port=7000)
         # self.client_2 = chromadb.HttpClient(host='10.6.82.192', port=8000)
 
     def __call__(self, *args, **kwargs):
@@ -51,10 +51,10 @@ class ChromaQuery(Blackbox):
             return JSONResponse(content={"error": "question is required"}, status_code=status.HTTP_400_BAD_REQUEST)
 
         if chroma_embedding_model is None or chroma_embedding_model.isspace() or chroma_embedding_model == "":
-            chroma_embedding_model = "/home/administrator/Workspace/Models/BAAI/bge-large-zh-v1.5"
+            chroma_embedding_model = "/home/gpu/Workspace/Models/BAAI/bge-large-zh-v1.5"
 
         if chroma_host is None or chroma_host.isspace() or chroma_host == "":
-            chroma_host = "172.16.5.8"
+            chroma_host = "10.6.81.119"
         
         if chroma_port is None or chroma_port.isspace() or chroma_port == "":
             chroma_port = "7000"
@@ -66,17 +66,17 @@ class ChromaQuery(Blackbox):
             chroma_n_results = 3
 
         # load client and embedding model from init
-        if re.search(r"172.16.5.8", chroma_host) and re.search(r"7000", chroma_port):
+        if re.search(r"10.6.81.119", chroma_host) and re.search(r"7000", chroma_port):
             client = self.client_1
         else:
             client = chromadb.HttpClient(host=chroma_host, port=chroma_port)
 
-        if re.search(r"/home/administrator/Workspace/Models/BAAI/bge-large-zh-v1.5", chroma_embedding_model):
+        if re.search(r"/home/gpu/Workspace/Models/BAAI/bge-large-zh-v1.5", chroma_embedding_model):
             embedding_model = self.embedding_model_1
-        elif re.search(r"/home/administrator/Workspace/Models/BAAI/bge-large-en-v1.5", chroma_embedding_model):
+        elif re.search(r"/home/gpu/Workspace/Models/BAAI/bge-small-en-v1.5", chroma_embedding_model):
             embedding_model = self.embedding_model_2
         else:
-            embedding_model = embedding_functions.SentenceTransformerEmbeddingFunction(model_name=chroma_embedding_model, device = "cuda")
+            embedding_model = embedding_functions.SentenceTransformerEmbeddingFunction(model_name=chroma_embedding_model, device = "cuda:1")
 
         # load collection
         collection = client.get_collection(chroma_collection_id, embedding_function=embedding_model)

src/blackbox/g2e.py

@@ -23,7 +23,7 @@ class G2E(Blackbox):
         if context == None:
             context = []
         #url = 'http://120.196.116.194:48890/v1'
-        url = 'http://120.196.116.194:48892/v1'
+        url = 'http://10.6.81.119:23333/v1'
 
         background_prompt = '''KOMBUKIKI是一款茶饮料，目标受众 年龄：20-35岁 性别：女性 地点：一线城市、二线城市 职业：精英中产、都市白领 收入水平：中高收入，有一定消费能力 兴趣和爱好：注重健康，有运动习惯
 
@@ -73,11 +73,11 @@ class G2E(Blackbox):
         response = client.chat.completions.create(
         model=model_name,
         messages=messages,
-            temperature=0.8,
-            top_p=0.8,
-            frequency_penalty=0.5,
-            presence_penalty=0.8,
-            stop=100
+            temperature="0.8",
+            top_p="0.8",
+            frequency_penalty="0.5",
+            presence_penalty="0.8",
+            stop="100"
         )
 
         fastchat_content = response.choices[0].message.content

src/blackbox/tts.py

@@ -2,27 +2,154 @@ import io
 import time
 from ntpath import join
 
+import requests
 from fastapi import Request, Response, status
 from fastapi.responses import JSONResponse
 from .blackbox import Blackbox
 from ..tts.tts_service import TTService
+from ..configuration import MeloConf
+from ..configuration import CosyVoiceConf
+from injector import inject
 from injector import singleton
 
+import sys,os
+sys.path.append('/home/gpu/Workspace/CosyVoice')
+from cosyvoice.cli.cosyvoice import CosyVoice
+from cosyvoice.utils.file_utils import load_wav
+
+import soundfile
+import pyloudnorm as pyln
+from melo.api import TTS as MELOTTS
+
+from ..log.logging_time import logging_time
+import logging
+logger = logging.getLogger(__name__)
+
 @singleton
 class TTS(Blackbox):
+    melo_mode: str
+    melo_url: str
+    melo_speed: int
+    melo_device: str
+    melo_language: str
+    melo_speaker: str
+
+    cosyvoice_mode: str
+    cosyvoice_url: str
+    cosyvoice_speed: int
+    cosyvoice_device: str
+    cosyvoice_language: str
+    cosyvoice_speaker: str
+
+    @logging_time(logger=logger)
+    def melo_model_init(self, melo_config: MeloConf) -> None:
+        self.melo_speed = melo_config.speed
+        self.melo_device = melo_config.device
+        self.melo_language = melo_config.language
+        self.melo_speaker = melo_config.speaker
+        self.melo_url = ''
+        self.melo_mode = melo_config.mode
+        self.melotts = None
+        self.speaker_ids = None
+        if self.melo_mode == 'local':
+            self.melotts = MELOTTS(language=self.melo_language, device=self.melo_device)
+            self.speaker_ids = self.melotts.hps.data.spk2id
+        else:
+            self.melo_url = melo_config.url
+        logging.info('#### Initializing MeloTTS Service in ' + self.melo_device + ' mode...')
+    
+    @logging_time(logger=logger)
+    def cosyvoice_model_init(self, cosyvoice_config: CosyVoiceConf) -> None:
+        self.cosyvoice_speed = cosyvoice_config.speed
+        self.cosyvoice_device = cosyvoice_config.device
+        self.cosyvoice_language = cosyvoice_config.language
+        self.cosyvoice_speaker = cosyvoice_config.speaker
+        self.cosyvoice_url = ''
+        self.cosyvoice_mode = cosyvoice_config.mode
+        self.cosyvoicetts = None
+        # os.environ['CUDA_VISIBLE_DEVICES'] = str(cosyvoice_config.device)
+        if self.cosyvoice_mode == 'local':
+            self.cosyvoicetts = CosyVoice('/home/gpu/Workspace/Models/CosyVoice/pretrained_models/CosyVoice-300M')
+            
+        else:
+            self.cosyvoice_url = cosyvoice_config.url
+        logging.info('#### Initializing CosyVoiceTTS Service in cuda:' + self.cosyvoice_device + ' mode...')
+
+    @inject   
+    def __init__(self, melo_config: MeloConf, cosyvoice_config: CosyVoiceConf, settings: dict) -> None:
+        self.tts_service = TTService("yunfeineo")
+        self.melo_model_init(melo_config)
+        self.cosyvoice_model_init(cosyvoice_config)
 
-    def __init__(self, *args, **kwargs) -> None:
-        self.tts_service = TTService("catmaid")
 
     def __call__(self, *args, **kwargs):
         return self.processing(*args, **kwargs)
 
-    def processing(self, *args, **kwargs) -> io.BytesIO:
+    @logging_time(logger=logger)
+    def processing(self, *args, settings: dict) -> io.BytesIO:
+
+        print("\nChat Settings: ", settings)
+        if settings is None:
+            settings = {}
+        user_model_name = settings.get("tts_model_name")
+        print(f"tts_model_name: {user_model_name}")
+
         text = args[0]
         current_time = time.time()
-        audio = self.tts_service.read(text)
-        print("#### TTS Service consume : ", (time.time()-current_time))
-        return audio
+
+        if user_model_name == 'melotts': 
+            if self.melo_mode == 'local':
+                audio = self.melotts.tts_to_file(text, self.speaker_ids[self.melo_speaker], speed=self.melo_speed)
+                f = io.BytesIO()
+                soundfile.write(f, audio, 44100, format='wav')     
+                f.seek(0)
+
+                # Read the audio data from the buffer
+                data, rate = soundfile.read(f, dtype='float32')
+                
+                # Peak normalization
+                peak_normalized_audio = pyln.normalize.peak(data, -1.0)
+                
+                # Integrated loudness normalization
+                meter = pyln.Meter(rate)
+                loudness = meter.integrated_loudness(peak_normalized_audio)
+                loudness_normalized_audio = pyln.normalize.loudness(peak_normalized_audio, loudness, -12.0)
+                
+                # Write the loudness normalized audio to an in-memory buffer
+                normalized_audio_buffer = io.BytesIO()
+                soundfile.write(normalized_audio_buffer, loudness_normalized_audio, rate, format='wav')
+                normalized_audio_buffer.seek(0)
+                
+                print("#### MeloTTS Service consume - local : ", (time.time() - current_time))
+                return normalized_audio_buffer.read()   
+
+            else:
+                message = {
+                    "text": text
+                }
+                response = requests.post(self.melo_url, json=message)
+                print("#### MeloTTS Service consume - docker : ", (time.time()-current_time))
+                return response.content 
+            
+        elif user_model_name == 'cosyvoicetts':
+            if self.cosyvoice_mode == 'local':
+                audio = self.cosyvoicetts.inference_sft(text, self.cosyvoice_language)
+                f = io.BytesIO()
+                soundfile.write(f, audio['tts_speech'].cpu().numpy().squeeze(0), 22050, format='wav')
+                f.seek(0)
+                print("#### CosyVoiceTTS Service consume - local : ", (time.time() - current_time))
+                return f.read()
+            else:
+                message = {
+                    "text": text
+                }
+                response = requests.post(self.cosyvoice_url, json=message)
+                print("#### CosyVoiceTTS Service consume - docker : ", (time.time()-current_time))
+                return response.content            
+        else:           
+            audio = self.tts_service.read(text)
+            print("#### TTS Service consume : ", (time.time()-current_time))
+            return audio.read()
 
     def valid(self, *args, **kwargs) -> bool:
         text = args[0]
@@ -31,10 +158,12 @@ class TTS(Blackbox):
     async def fast_api_handler(self, request: Request) -> Response:
         try:
             data = await request.json()
+            print(f"data: {data}")
         except:
             return JSONResponse(content={"error": "json parse error"}, status_code=status.HTTP_400_BAD_REQUEST)
         text = data.get("text")
+        setting = data.get("settings")
         if text is None:
             return JSONResponse(content={"error": "text is required"}, status_code=status.HTTP_400_BAD_REQUEST)
-        by = self.processing(text)
-        return Response(content=by.read(), media_type="audio/wav", headers={"Content-Disposition": "attachment; filename=audio.wav"})
+        by = self.processing(text, settings=setting)
+        return Response(content=by, media_type="audio/wav", headers={"Content-Disposition": "attachment; filename=audio.wav"})

src/blackbox/tts_bak.py

@@ -0,0 +1,40 @@
+import io
+import time
+from ntpath import join
+
+from fastapi import Request, Response, status
+from fastapi.responses import JSONResponse
+from .blackbox import Blackbox
+from ..tts.tts_service import TTService
+from injector import singleton
+
+@singleton
+class TTS(Blackbox):
+
+    def __init__(self, *args, **kwargs) -> None:
+        self.tts_service = TTService("yunfeineo")
+
+    def __call__(self, *args, **kwargs):
+        return self.processing(*args, **kwargs)
+
+    def processing(self, *args, **kwargs) -> io.BytesIO:
+        text = args[0]
+        current_time = time.time()
+        audio = self.tts_service.read(text)
+        print("#### TTS Service consume : ", (time.time()-current_time))
+        return audio
+
+    def valid(self, *args, **kwargs) -> bool:
+        text = args[0]
+        return isinstance(text, str)
+
+    async def fast_api_handler(self, request: Request) -> Response:
+        try:
+            data = await request.json()
+        except:
+            return JSONResponse(content={"error": "json parse error"}, status_code=status.HTTP_400_BAD_REQUEST)
+        text = data.get("text")
+        if text is None:
+            return JSONResponse(content={"error": "text is required"}, status_code=status.HTTP_400_BAD_REQUEST)
+        by = self.processing(text)
+        return Response(content=by.read(), media_type="audio/wav", headers={"Content-Disposition": "attachment; filename=audio.wav"})

src/blackbox/vlms.py

@@ -13,6 +13,9 @@ import base64
 import copy
 import ast
 
+import io 
+from PIL import Image
+from lmdeploy.serve.openai.api_client import APIClient
 import io 
 from PIL import Image
 from lmdeploy.serve.openai.api_client import APIClient
@@ -24,6 +27,7 @@ def is_base64(value) -> bool:
     except Exception:
         return False
 
+@singleton
 @singleton
 class VLMS(Blackbox):
 

src/configuration.py

@@ -1,138 +1,156 @@
-
-from dataclasses import dataclass
-from injector import inject,singleton
-import yaml
-import sys
-import logging
-
-@singleton
-class Configuration():
-
-    @inject
-    def __init__(self) -> None:
-        config_file_path = ""
-        try:
-            config_file_path = sys.argv[1]
-        except:
-            config_file_path = ".env.yaml"
-        with open(config_file_path) as f:
-            cfg = yaml.load(f, Loader=yaml.FullLoader)
-            self.cfg = cfg
-    
-    def getDict(self):
-        return self.cfg
-
-    """
-    # yaml 檔中的路徑 get("aaa.bbb.ccc")
-    aaa:
-      bbb:
-         ccc: "hello world"
-    """
-    def get(self, path: str | list[str], cfg: dict = None, default=None):
-        if isinstance(path, str):
-            if cfg is None:
-                cfg = self.cfg
-            return self.get(path.split("."), cfg)
-        length = len(path)
-        if length == 0  or not isinstance(cfg, dict):
-            return default
-        if length == 1:
-            return cfg.get(path[0])
-        return self.get(path[1:], cfg.get(path[0]))
-
-class TesouConf():
-    url: str
-    
-    @inject
-    def __init__(self, config: Configuration) -> None:
-        self.url = config.get("tesou.url")
-
-
-class MeloConf():
-    mode: str
-    url: str
-    speed: int
-    device: str
-    language: str
-    speaker: str
-
-    @inject
-    def __init__(self, config: Configuration) -> None:
-        self.mode = config.get("melotts.mode")
-        self.url = config.get("melotts.url")
-        self.speed = config.get("melotts.speed")
-        self.device = config.get("melotts.device")
-        self.language = config.get("melotts.language")
-        self.speaker = config.get("melotts.speaker")
-
-class CosyVoiceConf():
-    mode: str
-    url: str
-    speed: int
-    device: str
-    language: str
-    speaker: str
-
-    @inject
-    def __init__(self, config: Configuration) -> None:
-        self.mode = config.get("cosyvoicetts.mode")
-        self.url = config.get("cosyvoicetts.url")
-        self.speed = config.get("cosyvoicetts.speed")
-        self.device = config.get("cosyvoicetts.device")
-        self.language = config.get("cosyvoicetts.language")
-        self.speaker = config.get("cosyvoicetts.speaker")
-        
-# 'CRITICAL': CRITICAL,
-# 'FATAL': FATAL,
-# 'ERROR': ERROR,
-# 'WARN': WARNING,
-# 'WARNING': WARNING,
-# 'INFO': INFO,
-# 'DEBUG': DEBUG,
-# 'NOTSET': NOTSET,
-DEFAULT_LEVEL="WARNING"
-DEFAULT_TIME_FORMAT="%Y-%m-%d %H:%M:%S"
-
-@singleton
-class LogConf():
-    level: int
-    time_format = "%Y-%m-%d %H:%M:%S"
-    filename: str | None
-    @inject
-    def __init__(self, config: Configuration) -> None:
-        self.level = config.get("log.level")
-        c = config.get("log.level", default=DEFAULT_LEVEL).upper()
-        level=logging._nameToLevel.get(c)
-        if level is None:
-            self.level = logging.WARNING
-        else:
-            self.level = level
-        self.filename = config.get("log.filename")
-        self.time_format = config.get("log.time_format", default=DEFAULT_TIME_FORMAT)
-
-@singleton
-class EnvConf():
-    version: str
-    host: str
-    port: str
-    @inject
-    def __init__(self, config: Configuration) -> None:
-        self.version = "0.0.1"
-        self.host = config.get("env.host", default="0.0.0.0")
-        self.port = config.get("env.port", default="8080")
-
-@singleton
-@dataclass
-class BlackboxConf():
-    lazyloading: bool
-
-    @inject
-    def __init__(self, config: Configuration) -> None:
-        self.lazyloading = bool(config.get("blackbox.lazyloading", default=False))
-
-@singleton
-class VLMConf():
-
-    @inject
-    def __init__(self, config: Configuration) -> None:
-        self.url = config.get("vlms.url")
+
+from dataclasses import dataclass
+from injector import inject,singleton
+import yaml
+import sys
+import logging
+
+@singleton
+class Configuration():
+
+    @inject
+    def __init__(self) -> None:
+        config_file_path = ""
+        try:
+            config_file_path = sys.argv[1]
+        except:
+            config_file_path = ".env.yaml"
+        with open(config_file_path) as f:
+            cfg = yaml.load(f, Loader=yaml.FullLoader)
+            self.cfg = cfg
+    
+    def getDict(self):
+        return self.cfg
+
+    """
+    # yaml 檔中的路徑 get("aaa.bbb.ccc")
+    aaa:
+      bbb:
+         ccc: "hello world"
+    """
+    def get(self, path: str | list[str], cfg: dict = None, default=None):
+        if isinstance(path, str):
+            if cfg is None:
+                cfg = self.cfg
+            return self.get(path.split("."), cfg)
+        length = len(path)
+        if length == 0  or not isinstance(cfg, dict):
+            return default
+        if length == 1:
+            return cfg.get(path[0])
+        return self.get(path[1:], cfg.get(path[0]))
+
+class TesouConf():
+    url: str
+    
+    @inject
+    def __init__(self, config: Configuration) -> None:
+        self.url = config.get("tesou.url")
+
+
+class MeloConf():
+    mode: str
+    url: str
+    speed: int
+    device: str
+    language: str
+    speaker: str
+
+    @inject
+    def __init__(self, config: Configuration) -> None:
+        self.mode = config.get("melotts.mode")
+        self.url = config.get("melotts.url")
+        self.speed = config.get("melotts.speed")
+        self.device = config.get("melotts.device")
+        self.language = config.get("melotts.language")
+        self.speaker = config.get("melotts.speaker")
+
+class CosyVoiceConf():
+    mode: str
+    url: str
+    speed: int
+    device: str
+    language: str
+    speaker: str
+
+    @inject
+    def __init__(self, config: Configuration) -> None:
+        self.mode = config.get("cosyvoicetts.mode")
+        self.url = config.get("cosyvoicetts.url")
+        self.speed = config.get("cosyvoicetts.speed")
+        self.device = config.get("cosyvoicetts.device")
+        self.language = config.get("cosyvoicetts.language")
+        self.speaker = config.get("cosyvoicetts.speaker")
+
+class SenseVoiceConf():
+    mode: str
+    url: str
+    speed: int
+    device: str
+    language: str
+    speaker: str
+
+    @inject
+    def __init__(self, config: Configuration) -> None:
+        self.mode = config.get("sensevoiceasr.mode")
+        self.url = config.get("sensevoiceasr.url")
+        self.speed = config.get("sensevoiceasr.speed")
+        self.device = config.get("sensevoiceasr.device")
+        self.language = config.get("sensevoiceasr.language")
+        self.speaker = config.get("sensevoiceasr.speaker")        
+        
+# 'CRITICAL': CRITICAL,
+# 'FATAL': FATAL,
+# 'ERROR': ERROR,
+# 'WARN': WARNING,
+# 'WARNING': WARNING,
+# 'INFO': INFO,
+# 'DEBUG': DEBUG,
+# 'NOTSET': NOTSET,
+DEFAULT_LEVEL="WARNING"
+DEFAULT_TIME_FORMAT="%Y-%m-%d %H:%M:%S"
+
+@singleton
+class LogConf():
+    level: int
+    time_format = "%Y-%m-%d %H:%M:%S"
+    filename: str | None
+    @inject
+    def __init__(self, config: Configuration) -> None:
+        self.level = config.get("log.level")
+        c = config.get("log.level", default=DEFAULT_LEVEL).upper()
+        level=logging._nameToLevel.get(c)
+        if level is None:
+            self.level = logging.WARNING
+        else:
+            self.level = level
+        self.filename = config.get("log.filename")
+        self.time_format = config.get("log.time_format", default=DEFAULT_TIME_FORMAT)
+
+
+@singleton
+class EnvConf():
+    version: str
+    host: str
+    port: str
+    @inject
+    def __init__(self, config: Configuration) -> None:
+        self.version = "0.0.1"
+        self.host = config.get("env.host", default="0.0.0.0")
+        self.port = config.get("env.port", default="8080")
+
+@singleton
+@dataclass
+class BlackboxConf():
+    lazyloading: bool
+
+    @inject
+    def __init__(self, config: Configuration) -> None:
+        self.lazyloading = bool(config.get("blackbox.lazyloading", default=False))
+
+@singleton
+class VLMConf():
+
+    @inject
+    def __init__(self, config: Configuration) -> None:
+        self.url = config.get("vlms.url")

src/tts/tts_service.py

@@ -21,6 +21,60 @@ logging.basicConfig(level=logging.INFO)
 dirbaspath = __file__.split("\\")[1:-1]
 dirbaspath= "/home/gpu/Workspace/jarvis-models/src/tts" + "/".join(dirbaspath)
 config = {
+    'ayaka': {
+        'cfg': dirbaspath + '/models/ayaka.json',
+        'model': dirbaspath + '/models/ayaka_167k.pth',
+        'char': 'character_ayaka',
+        'speed': 1
+    },
+    'catmix': {
+        'cfg': dirbaspath + '/models/catmix.json',
+        'model': dirbaspath + '/models/catmix_107k.pth',
+        'char': 'character_catmix',
+        'speed': 1.1
+    },
+    'noelle': {
+        'cfg': dirbaspath + '/models/noelle.json',
+        'model': dirbaspath + '/models/noelle_337k.pth',
+        'char': 'character_noelle',
+        'speed': 1.1
+    },
+    'miko': {
+        'cfg': dirbaspath + '/models/miko.json',
+        'model': dirbaspath + '/models/miko_139k.pth',
+        'char': 'character_miko',
+        'speed': 1.1
+    },
+    'nahida': {
+        'cfg': dirbaspath + '/models/nahida.json',
+        'model': dirbaspath + '/models/nahida_129k.pth',
+        'char': 'character_nahida',
+        'speed': 1.1
+    },    
+    'ningguang': {
+        'cfg': dirbaspath + '/models/ningguang.json',
+        'model': dirbaspath + '/models/ningguang_179k.pth',
+        'char': 'character_ningguang',
+        'speed': 1.1
+    },    
+    'yoimiya': {
+        'cfg': dirbaspath + '/models/yoimiya.json',
+        'model': dirbaspath + '/models/yoimiya_102k.pth',
+        'char': 'character_yoimiya',
+        'speed': 1.1
+    },    
+    'yunfeineo': {
+        'cfg': dirbaspath + '/models/yunfeineo.json',
+        'model': dirbaspath + '/models/yunfeineo_25k.pth',
+        'char': 'character_yunfeineo',
+        'speed': 1.1
+    },    
+    'zhongli': {
+        'cfg': dirbaspath + '/models/zhongli.json',
+        'model': dirbaspath + '/models/zhongli_44k.pth',
+        'char': 'character_',
+        'speed': 1.1
+    },     
     'paimon': {
         'cfg': dirbaspath + '/models/paimon6k.json',
         'model': dirbaspath + '/models/paimon6k_390k.pth',
@@ -28,7 +82,7 @@ config = {
         'speed': 1
     },
     'yunfei': {
-        'cfg': dirbaspath + '/tts/models/yunfeimix2.json',
+        'cfg': dirbaspath + '/models/yunfeimix2.json',
         'model': dirbaspath + '/models/yunfeimix2_53k.pth',
         'char': 'character_yunfei',
         'speed': 1.1